Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

I have seen the following formula in papers (without citations) and in Mathematica's documentation about Tanh[]:

$$ \frac{\tanh(z)}{8z}=\sum_{k=1}^{\infty} \frac{1}{(2k-1)^2 \pi^2+4z^2} $$

I have no idea how to prove it and I have also encountered in my research similar sums involving, for instance, $\mathrm{coth}$. It would be nice to have a general method for working with these problems; any suggestions?

share|improve this question
    
Maybe you can try the Taylor Series? –  ᴊ ᴀ s ᴏ ɴ Jul 27 '12 at 14:14
1  
As noted, you should consult a textbook in complex analysis. Usually you will see this as a series for $\tan z$, and then a simple change of variables $iz$ will give you a similar one for $\tanh z$. –  GEdgar Jul 28 '12 at 12:36

3 Answers 3

up vote 8 down vote accepted

There is the infinite product representation

$$\cosh\,z=\prod_{k=1}^\infty \left(1+\frac{4z^2}{\pi^2(2k-1)^2}\right)$$

Taking logarithms gives

$$\log\cosh\,z=\sum_{k=1}^\infty \log\left(1+\frac{4z^2}{\pi^2(2k-1)^2}\right)$$

If we differentiate both sides, we have

$$\tanh\,z=\sum_{k=1}^\infty \frac{\frac{8z}{\pi^2(2k-1)^2}}{1+\frac{4z^2}{\pi^2(2k-1)^2}}$$

which simplifies to

$$\tanh\,z=\sum_{k=1}^\infty \frac{8z}{4z^2+\pi^2(2k-1)^2}$$

Note that the infinite product that we started with is the factorization of $\cosh$ over its (imaginary) zeroes.

Here is a related question.

share|improve this answer
    
@jm this is great; does all such formulas rely on manipulating other known products and sums? Isn't there, for instance, a Fourier transform or generating function approach (or something...)? Again, thank you for the support. –  Gabriel Landi Jul 28 '12 at 1:29
    
Well, the more systematic route uses polygamma functions/generalized harmonic numbers, but I thought you'd prefer the route that doesn't require special functions. This is the easiest route I know for the tangent, cotangent, and their hyperbolic counterparts. –  J. M. Jul 28 '12 at 1:35
    
Yeah. Probably. Hehe. It's just that I never know how to deal with sums in any other way than comparing it to other sums. Again, thank you for the answer J.M. –  Gabriel Landi Jul 28 '12 at 1:41
    
Our answers to this question seem to be paralleled here :-) –  robjohn Jul 28 '12 at 13:23
    
@rob, hence the upvote. ;) –  J. M. Jul 28 '12 at 13:49

In this answer, it is shown that for all $z\in\mathbb{C}\setminus\mathbb{Z}$, $$ \pi\cot(\pi z)=\frac1z+\sum_{k=1}^\infty\frac{2z}{z^2-k^2}\tag{1} $$ Applying the identity $\tan(x)=\cot(x)-2\cot(2x)$ to $(1)$ gives $$ \begin{align} \pi\tan(\pi z) &=\sum_{k=1}^\infty\frac{2z}{z^2-k^2}-\sum_{k=1}^\infty\frac{2z}{z^2-\frac{k^2}{4}}\\ &=\sum_{k=1}^\infty\frac{8z}{(2k-1)^2-4z^2}\tag{2} \end{align} $$ Applying the identity $\tanh(x)=-i\tan(ix)$ to $(2)$ yields $$ \pi\tanh(\pi z)=\sum_{k=1}^\infty\frac{8z}{(2k-1)^2+4z^2}\tag{3} $$ Finally, applying the change variables $z\mapsto z/\pi$ to $(3)$ reveals $$ \frac{\tanh(z)}{8z}=\sum_{k=1}^\infty\frac{1}{(2k-1)^2\pi^2+4z^2}\tag{4} $$

share|improve this answer

A general approach to find such expansions is a technique based on expanding a meromorphic function in terms of rational fractions by exploiting its poles. A theorem related to this work is called the Mittag-Leffler theorem in complex analysis. See here (starting from page (96)). You can find examples of such expansions for the functions $\tan(z)\,,\sec(z)\, \cot(z)\,, \csc(z) \,. $

share|improve this answer
4  
For a somewhat less obscene reference on Mittag-Leffler theorem, see here. –  Did Aug 24 '12 at 11:43
    
I reffered the OP where he can find the techniques and examples to his question. It is in the heart of my research. –  Mhenni Benghorbal Sep 11 '12 at 22:35

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.